demobook

Maintain visual consistency with Google Flow's 'Add to Prompt'

Demo summary

The user demonstrates a workflow in Google Flow where each generated scene is added back into the prompt to lock in lighting, style, and character identity for the next scene.

Step-by-step

  1. Copy the first scene text-to-image prompt from Claude
  2. Paste the prompt into Google Flow and hit generate
  3. Click the three dots in the top right corner of the generated image
  4. Select 'Add to prompt' to lock in visual consistency
  5. Copy the next scene prompt from Claude and paste it into Flow
  6. Repeat the 'Add to prompt' step for each new image before generating the subsequent scene

Watch out for

  • The prompt itself must be structured for consistency with locked characters and styles for this method to work

Tips

  • Ensure your prompt starts with specific 'locked' identifiers (e.g., 'use locked boy') to keep characters consistent
  • Add each image to the prompt before generating the next so each new scene starts from the previous one
  • Use this workflow to maintain a cohesive cinematic feel and color palette across a series of images

Highlights

When your prompt is built right and you add each image to the prompt before generating the next, that's when the magic happens.

All demos from “These 2 FREE Google AI Tools Just Changed 3D Animation Forever (Veo 3.1)

  1. 2:250:43Generate consistent characters in Google FlowThe video shows how to use Google Flow to generate a single reference image containing all main characters to ensure visual consistency across the film.Google FlowAI Person Generator
  2. 3:212:08Maintain visual consistency with Google Flow's 'Add to Prompt'CurrentThe user demonstrates a workflow in Google Flow where each generated scene is added back into the prompt to lock in lighting, style, and character identity for the next scene.Google FlowAI Image Generator
  3. 6:141:06Animate images with Google Veo 3.1 in Google VidsThe creator demonstrates uploading a still image to Google Vids and using the Veo 3.1 model to generate an animated clip with automatic lip-sync and sound effects.Google VidsImage to Video
  4. 9:360:24Add automatic captions in Google VidsThe video shows the one-click captioning feature in Google Vids that automatically transcribes and syncs dialogue from the generated AI video clips.Google VidsAI Video Editor
  5. Watch “These 2 FREE Google AI Tools Just Changed 3D Animation Forever (Veo 3.1)” →

AI Image Generator

  1. 2:061:23Generate a character reference sheet in Google FlowThe creator uploads a Midjourney image to Google Flow and uses a specific layout prompt with the Nano Banana model to generate a multi-angle character reference sheet, then demonstrates how to upscale the result to 2K.AI Samson
  2. 3:212:08Maintain visual consistency with Google Flow's 'Add to Prompt'CurrentThe user demonstrates a workflow in Google Flow where each generated scene is added back into the prompt to lock in lighting, style, and character identity for the next scene.The AI Blueprint
  3. 1:370:38Configuring image generation in Google Labs FlowThe user demonstrates setting up a project in Flow by Google Labs, switching the mode to 'create image' and selecting the Nano Banana Pro model with portrait orientation.Ryan Hogue Passive Income
  4. 4:400:18Generating AI graphics with Nano Banana ProThe video shows the output of four AI-generated graphics in Flow, demonstrating the grid view and the quality of the Nano Banana Pro model results.Ryan Hogue Passive Income
  5. 2:010:21Generate scene-specific images in FlowThe narrator demonstrates how to use scene prompts and the character reference feature (plus button) in Flow to maintain consistency across generated scene images.Autotube Academy
  6. 3:410:28Create consistent scenes using image referencesThe narrator demonstrates using the reference button in Flow to upload character images, ensuring visual consistency across generated scene backgrounds.EarnWithAi
  7. 16:120:31Comparing Fast vs Quality modelsThe user navigates the settings to switch from the fast model to the Veo 3.1 quality version to demonstrate the difference in credit cost and output results.AI2Play